Support Vector Machine을 이용한 냉간 압연기의 판 터짐 결함 검지

Abstract

There are several types of fault occurred in cold rolling mill. In case of common faults such as strip breakage and chattering have been studied by some researchers. Furthermore, the study of strip rupture has been developing due to difficulties of signal analysis. The main reason is that fault usually happens in transient section. In this paper, we propose a fault detection system to detect the strip rupture in six-high stand cold rolling mills based on transient current signal of an electrical motor. In this work, signal smoothing technique is used to highlight feature precisely between normal and fault condition. Subtracting the smoothed signal from the original signal gives the residuals that contain the information related to the normal or faulty condition. Then this signal is segmented to get more information from original signal. Using segmented residual signal, discrete wavelet transform (DWT) is performed and acquired the signal presenting fault feature well. Also, feature extraction and classification are employed by using principle component analysis (PCA), independent component analysis (ICA), kernel principle component analysis (KPCA), kernel independent component analysis (KICA), and support vector machine (SVM). The actual data were acquired from the domestic steel company for validating the proposed method.